IgM,but not IgA rheumatoid factor interferes with anti-cardiolipin and anti-β2 glycoprotein I measurements: a quantitative analysis

Abstract

IgM rheumatoid factor (RF) is sometimes referred to as capable of causing interference in the IgM anti-cardiolipin (aCL) testing. Published guidelines are, however, inconsistent, and evidence regarding the interference is limited. Our goal was investigate IgM and IgA RF cross-reactivity and/or interference in IgM and IgA aCL and anti-β2 glycoprotein I (aβ2GPI) testing. Serum specimens with high IgM and IgA RF levels were tested for IgG, IgA and IgM aCL and aβ2GPI antibodies to examine cross-reactivity. Samples containing IgG aCL and aβ2GPI antibodies were spiked with IgM (and IgA) RF, and samples with high RF levels were spiked with IgG aCL antibodies. The mixtures were tested for IgM and IgA aCL and aβ2GPI antibodies. Specimens with high IgM and IgA RF concentrations did not test positive for IgM or IgA aCL and aβ2GPI antibodies (except one weak positive IgA aβ2GPI result), indicating the lack of cross-reactivity. In the spiked specimens, addition of IgM RF caused significant positive bias in the measurement of both aCL and aβ2GPI antibodies of IgM isotype in the presence of IgG aCL and aβ2GPI antibodies. The threshold for triggering significant interference was 318IU/ml for IgM RF, and 77 GPLU/ml for IgG aCL. Neither IgM, nor IgA RF, however, affected the IgA antiphospholipid (aPL) antibody testing. IgM RF can cause a false-positive IgM aCL result in the presence of IgG aCL antibodies. In studies on the prevalence and clinical significance of IgM aPL antibodies, RF interference should be considered and RF testing should be performed.

Keywords

anti-cardiolipin antibody IgA IgM interference rheumatoid factor

Introduction

Antiphospholipid antibody (aPL) measurements are routinely used in the laboratory evaluation of patients with antiphospholipid syndrome (APS). The most recent consensus guidelines recognize lupus anticoagulant (LA), anti-cardiolipin (aCL) and anti-β2 glycoprotein I (aβ2GPI) antibodies as appropriate laboratory criteria for classification of APS.¹ The presence of medium or high titer of aCL or aβ2GPI antibodies of either IgG or IgM isotypes is sufficient for the laboratory diagnosis of APS.¹ Several studies, however, have demonstrated the lower specificity of IgM aCL antibodies compared with that of the IgG isotype for the features of APS.²

IgM rheumatoid factor (RF) is sometimes mentioned as one of the factors capable of causing interference in the IgM aCL assays. Published aPL antibody testing guidelines are, however, inconsistent. The consensus guidelines on aCL and aβ2GPI testing and reporting by the Australasian Anticardiolipin Working Party and a recently published aCL ELISA protocol state that RF may produce false-positive IgM aCL results.³ ^– ⁵ On the other hand, neither the methodological recommendations proposed by the European Forum on Antiphospholipid Antibodies,⁶ nor the current consensus statement on the classification criteria¹ or recently published comprehensive reviews on the laboratory evaluation of APS^7,8 mention such an interference. Review of package inserts of 11 commercially available aCL immunoassays has revealed that five of them include a warning regarding RF interference in IgM aCL measurement; one of them states that no cross-reactivity was observed with RF; and five do not mention RF at all. For aβ2GPI kits, the numbers in the same categories are three, one and five out of nine. Even when interference is mentioned, the nature, degree or mechanism of the phenomenon is not described. Most frequently a general warning is issued, stating that RF may produce false-positive IgM aCL results. Although interference with hemoglobin, bilirubin, triglycerides and other known interfering factors is usually properly evaluated and quantified by manufacturers, the information on the concentration of RF capable of causing interference and the degree of potential bias is lacking. Laboratory personnel and clinicians do not have evidence-based information on how to interpret IgM aCL results when RF is present. There have been only two publications in the scientific literature on the topic; both are one-page Letters, and they contain only limited data. One report found that a subset of RFs cross-reacts with cardiolipin and can cause false-positive IgM aCL results.⁹ The other suggests that RF binds to IgG aCL (bound to the solid-phase antigen) and thereby causes the serum to appear positive for IgM aCL.¹⁰

Our goal was to obtain definitive proof for the presence or absence of IgM RF cross-reactivity and/or interference in IgM aCL and aβ2GPI measurements. We also aimed at quantifying the interference and determining the threshold RF concentration capable of causing bias. In addition, we wished to examine whether IgA RF could interfere with IgA aCL and aβ2GPI testing.

Methods

Cross-reactivity

To examine potential cross-reactivity, nine serum specimens were selected based on their high IgM RF positivity. They also contained IgA RF, and some of them IgG RF as well (Table 1). IgG, IgA and IgM aCL and aβ2GPI antibodies were measured in all serum samples by the EL-aCL™ and EL-aβ2GPI™ kits from TheraTest Laboratories, Inc. (Lombard, IL, USA), according to the manufacturer’s instructions.

Table 1

The prevalence and level of RF, aCL and aβ2GPI antibodies of IgM, IgA and IgG isotypes in the nine sera included in the cross-reactivity study

	RF			aCL			aβ2GPI
	IgM IU/ml	IgA U/ml	IgG U/ml	IgM MPLU/ml	IgA APLU/ml	IgG GPLU/ml	IgM U/ml	IgA U/ml	IgG U/ml
Serum #1	8616	502	59	5.9	0.7	2.2	0.9	2.3	9.2
Serum #2	4939	446	12	13.1*	0.6	1.4	5.3*	7.9	5.4
Serum #3	3276	484	34	8.4	0.3	0.7	2.9	2.0	3.8
Serum #4	913	436	136	9.2	0.7	2.1	3.7	2.8	6.7
Serum #5	4652	362	19	2.3	0.4	19.8*	0.8	3.9	10.0
Serum #6	476	401	60	0.9	0.6	1.9	0.3	4.0	10.1
Serum #7	812	158	46	2.4	0.2	1.8	1.0	1.7	146.7
Serum #8	946	113	69	1.9	0.3	1.1	1.0	5.3*	6.1
Serum #9	380	150	33	1.5	0.5	2.2	1.7	3.4	11.8

Positive results are in Bold.

Equivocal results.

Interference

To assess interference, serum samples containing various levels of IgG aCL and aβ2GPI antibodies (from negative to high positive) were mixed in 3:1, 1:1 and 1:3 ratios with specimens containing high levels of IgM and IgA RF. The original sera, as well as the mixtures, were tested with the TheraTest EL-aCL™ and EL-aβ2GPI ™ kits for IgM and IgA antibodies. The obtained results were compared with the expected antibody levels, calculated based on the original samples’ antibody concentration. The interference was considered significant if the bias was equal to or greater than the cut-off value of the particular test.

To determine the threshold RF concentration for interference, a specimen containing 177GPLU/ml IgG aCL antibodies was spiked with various amounts of IgM RF. IgM aCL levels were measured and the bias was calculated as described above.

To calculate the effect of IgG antibody concentration on the interference, various amounts of aCL IgG antibodies were added to a serum containing 473IU/ml IgM RF, and IgM aCL levels were then measured and the bias was calculated according to the method described above.

Correlations were assessed by regression analysis. Results were considered significant if the p-value was <0.05. For all experiments, de-identified leftover serum specimens were used from the reference laboratory of TheraTest Laboratories, Inc, as approved by Chesapeake Research Review, Inc. (Columbia, MD, USA), an independent research review board.

RF interference in patients with APS

To assess the potential occurrence of IgM RF interference in APS, 26 serum specimens from APS patients were assayed for IgG, IgA and IgM aCL and aβ2GPI antibodies with the TheraTest EL-aCL™ and EL- aβ2GPI ™ kits, and for IgM RF with the TheraTest EL-RF/3™ kit, according to the manufacturer’s instructions. The samples were selected to contain various levels and isotypes of aPL antibodies, and were kindly provided by Silvia S Pierangeli (Professor of Medicine, Microbiology and Pathology at UTMB and Director of Louisville APL Diagnostics, Inc.). These specimens were previously assayed during the wet workshop of the 13th International Congress on Antiphospholipid Antibodies, Galveston, Texas, USA, in April 2010.

Results

Cross-reactivity

Out of the nine high IgM and IgA RF positive sera, none was IgM aCL and aβ2GPI positive. Serum #2 tested equivocal for IgM aCL and IgM aβ2GPI antibodies. The same serum tested weakly positive for IgA aβ2GPI, and Serum #8 was equivocal (Table 1). Serum #7 was repeatedly highly positive for IgG aβ2GPI antibodies, but negative for IgG aCL. Interference caused by IgG RF was not subject of this study, but the low level of IgG RF and the lack of positivity in other IgG RF specimens imply true IgG aβ2GPI positivity.

Interference

When IgG aCL and aβ2GPI negative specimens were mixed with IgM (and IgA) RF-containing sera, the obtained IgM aCL and aβ2GPI antibody levels were not significantly different from the expected values. However, when specimens containing medium or high levels of IgG aCL and aβ2GPI antibodies were mixed with RF-containing samples, the obtained IgM antibody levels were substantially higher than the expected results. When the bias (the difference between the obtained and expected IgM antibody values) was plotted against the IgG aCL and aβ2GPI antibody level in the mixture, a significant positive correlation was revealed (r = 0.960, p < 0.001 for aCL; r = 0.870, p < 0.001 for aβ2GPI) (Figures 1 and 2).

When the same specimen mixtures were tested in the IgA aCL and aβ2GPI assays, the obtained IgA antibody levels were the same as the expected values, indicating the lack of interference. The maximum difference between the obtained and expected antibody levels was <6% of the cut-off value for IgA aCL antibodies, and <23% of the cut-off value for IgA aβ2GPI antibodies (data not shown).

Spiking of an aCL IgG-positive specimen (177GPLU/ml) with various concentrations of IgM RF resulted in significant positive bias in the aCL IgM measurement, which strongly correlated with the concentration of the added IgM RF (r = 0.987, p < 0.001; Figure 3). Based on the regression equation, the threshold of IgM RF for triggering significant (≥cut-off value of the IgM aCL test) interference was calculated as 318IU/ml.

When different amounts of aCL IgG antibodies were added to a serum containing 473IU/ml IgM RF, the bias of the IgM aCL measurements showed strong positive correlation with the amount of added aCL IgG (r = 0.997, p < 0.001; Figure 4). Based on the regression equation, 77GPLU/ml aCL antibody in the sample was able to cause significant (≥cut-off value) interference in the aCL IgM measurement. It is important to note, however, that the bias is measured on a continuous scale, and it is detectable even below this aCL IgG level.

The occurrence of RF interference in patients with APS

Of the 26 APS specimens, nine were IgG aCL and IgM aCL double positive (with or without IgA antibodies), 15 were IgG aCL positive only, and two were IgM aCL positive only. None was IgM RF positive.

IgM RF at 473IU/ml was added to two of the APS specimens containing 179 and 361GPLU/ml IgG aCL antibodies (final concentrations after spiking), but negative for IgM aCL (1 and 3MPLU/ml, respectively). The IgM aCL results of the spiked specimens were 31 and 51MPLU/ml (cut-off: 10MPLU/ml).

Discussion

Our data confirm that IgM RF does interfere with IgM aCL and aβ2GPI measurements, and can cause false-positive results. Neither IgM, nor IgA RF, however, affect the IgA aPL antibody testing.

RF is one of the main sources of interference in immunometric (sandwich-type) immunoassays.¹¹ ^– ¹³ Its influence on autoantibody testing, however, has not been well established. According to our findings, IgM RF itself does not cross-react with the CL or β2GPI antigen, hence the presence of IgM RF alone does not convert a negative specimen to an apparent IgM aCL or IgM aβ2GPI-positive one. However, the co-existence of a medium or high level of IgG aCL/aβ2GPI antibody and a high titer of IgM RF in the specimen results in significant, dose-dependent positive bias in the IgM aCL and aβ2GPI measurement. The degree of interference depends on the concentration of both the IgM RF and the IgG aCL/aβ2GPI antibodies. In our system, 318IU/ml RF IgM was the threshold for triggering significant (≥cut-off value of the IgM aCL test) interference. One should always remember though, that a continuous, linear relationship exists between the IgM RF concentration and the extent of interference it causes; consequently, the degree of interference is measurable even below this value. On the other hand, 77GPLU/ml of IgG aCL antibody level was found to be sufficient to generate significant interference in the aCL IgM testing. According to the same principal, measurable (although less than the cut-off concentration) interference can still be detected below this level. Spiking of IgM aCL negative (but IgG aCL positive) specimens with IgM RF converted them to medium/highly IgM aCL positive, clearly indicating that IgM RF can cause false-positive IgM aCL results.

The consequences of these findings vary according to the purpose of the testing. Although IgM RF can clearly cause false-positive IgM aCL results in the presence of IgG aCL antibodies, our data suggest that in diagnostic situations the interference probably does not result in misclassification or misdiagnosis. The bias is significant only when the IgG aCL antibody level exceeds the medium titer, which alone is sufficient for the diagnosis of APS. The combination of a lower IgM RF and aCL IgG concentration, however, can still affect the accuracy of the results. It can be sufficient to generate minor shifts, and convert a negative result to equivocal or low positive, or push a weak positive value to the medium range.

When the purpose of testing is to study the prevalence of IgM aPL antibodies and seek correlations with clinical manifestations, RF interference should always be considered and RF testing should be performed in order to obtain reliable data. IgM aCL/aβ2GPI antibodies have been considered less useful for the diagnosis and management of APS, and the association between IgM antibodies and thrombosis is weaker than that of the IgG isotype.^2,14,15 It is possible that previously undetected RF interference has biased the data and contributed to these findings. Moreover, although our experiments were confined to aCL and aβ2GPI measurements only, it is reasonable to assume that other aPL antibody assays, including anti-prothrombin, anti-phosphatidylserine and other aPL antibody tests are similarly affected.

The mechanism of the interference is probably IgM RF binding to IgG antibodies bound to the solid-phase antigen, as has been previously demonstrated by Agopian et al.¹⁰ The lack of IgA RF interference might be explained by competition between IgM and IgA RF for the limited amount of antigen, and the possibly higher concentration and/or affinity and avidity of IgM RF compared with IgA RF. IgA aβ2GPI antibodies are often present in APS,¹⁶ ^– ¹⁸ and recently they have been implicated as lupus-specific antibodies. Our results, showing the lack of IgA RF interference, confirm the analytical specificity of IgA aCL and aβ2GPI measurements, and strengthen the significance of those findings.

None of the 26 APS patients with various levels of IgG and IgM aCL antibodies tested positive for IgM RF. Although this was only a small cohort of patients, the results suggest that RF positivity and consequent RF interference is probably not a major source of inaccurate results.

Every test system is different regarding the amount of solid-phase-bound antigen, serum dilution, incubation conditions, detection method, etc. Given the well-known variability of results obtained with different aPL antibody tests, interference data (especially the threshold levels for interference) generated by other aPL assays may be different from these results, obtained with TheraTest kits. Researchers and manufacturers of aPL antibody assays are encouraged to obtain data about their own specific test systems. Our findings, however, shed light on the general mechanism of RF interference in aPL antibody measurements, and help laboratory scientists and clinicians in the interpretation of aPL antibody test results.

Footnotes

Acknowledgment

We wish to thank Kathy L Russo for her excellent technical contribution, W Carey Hanly for useful comments and suggestions, and Silvia S Pierangeli for providing the APS serum samples.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest

G Lakos and M Teodorescu are employees of TheraTest Laboratories, Inc.

References

Miyakis

Lockshin

Atsumi

. International consensus statement on an update of the classification criteria for definite antiphospholipid syndrome (APS). J Thromb Haemost 2006; 4: 295–306.

Galli

Luciani

Bertolini

Barbui

. Lupus anticoagulants are stronger risk factors for thrombosis than anticardiolipin antibodies in the antiphospholipid syndrome: a systematic review of the literature. Blood 2003; 101: 1827–1832.

Wong

Gillis

Adelstein

. Consensus guidelines on anti-cardiolipin antibody testing and reporting. Pathology 2004; 36: 63–68.

Wong

Favaloro

Adelstein

. Consensus guidelines on anti-beta 2 glycoprotein I testing and reporting. Pathology 2008; 40: 58–63.

Pierangeli

Harris

. A protocol for determination of anticardiolipin antibodies by ELISA. Nat Protoc 2008; 3: 840–848.

Tincani

Allegri

Balestrieri

. Minimal requirements for antiphospholipid antibodies ELISAs proposed by the European Forum on Antiphospholipid Antibodies. Thromb Res 2004; 114: 553–558.

Abo

DeBari

. Laboratory evaluation of the antiphospholipid syndrome. Ann Clin Lab Sci 2007; 37: 3–14.

Devreese

Hoylaerts

. Challenges in the diagnosis of the antiphospholipid syndrome. Clin Chem 2010; 56: 930–940.

Galperin

Fortin

Subang

Newkirk

Rauch

. A subset of rheumatoid factors crossreacts with cardiolipin in patients positive for IgM rheumatoid factor and anticardiolipin antibodies. J Rheumatol 2000; 27: 820–821.

10.

Agopian

Boctor

Peter

. False-positive test result for IgM anticardiolipin antibody due to IgM rheumatoid factor. Arthritis Rheum 1988; 31: 1212–1213.

11.

Martín

Marquet

Ferrer

. Rheumatoid factor interference in a tacrolimus immunoassay. Ther Drug Monit 2009; 31: 743–745.

12.

Berth

Bosmans

Everaert

. Rheumatoid factor interference in the determination of carbohydrate antigen 19-9 (CA 19-9). Clin Chem Lab Med 2006; 44: 1137–1139.

13.

Després

Grant

. Antibody interference in thyroid assays: a potential for clinical misinformation. Clin Chem 1998; 44: 440–454.

14.

Danowski

de Azevedo

de Souza Papi

Petri

. Determinants of risk for venous and arterial thrombosis in primary antiphospholipid syndrome and in antiphospholipid syndrome with systemic lupus erythematosus. J Rheumatol 2009; 36: 1195–1199.

15.

Swadźba

Iwaniec

Szczeklik

Musiał

. Revised classification criteria for antiphospholipid syndrome and the thrombotic risk in patients with autoimmune diseases. J Thromb Haemost 2007; 5: 1883–1889.

16.

Fanopoulos

Teodorescu

Varga

Teodorescu

. High frequency of abnormal levels of IgA anti-beta2-glycoprotein I antibodies in patients with systemic lupus erythematosus: relationship with antiphospholipid syndrome. J Rheumatol 1998; 25: 675–680.

17.

Lakos

Kiss

Regëczy

. Isotype distribution and clinical relevance of anti-beta2-glycoprotein I (beta2-GPI) antibodies: Importance of IgA isotype. Clin Exp Immunol 1999; 117: 574–579.

18.

Tsutsumi

Matsuura

Ichikawa

Fujisaku

Mukai

Koike

. IgA class anti-beta2-glycoprotein I in patients with systemic lupus erythematosus. J Rheumatol 1998; 25: 74–78.